915 1500words

From the Treatise on Money to The General Theory: John Maynard Keynes’s Departure from the Doctrine of Forced Saving

Ho-Po Crystal Wong

John Maynard Keynes described his own General Theory of Employ- ment, Interest, and Money as “a long struggle of escape . . . from habitual modes of thought and expression. . . . The difficulty lies, not in the new ideas, but in escaping from the old ones, which ramify, for those brought up as most of us have been, into every corner of our minds” (JMK, 7:xxiii). There has been a substantial amount of scholarly interest in Keynes’s changing views on monetary theory (for instance, Patinkin 1976, 1982; Dimand 1988; Meltzer 1988; Amadeo 1989; Moggridge 1992; Skidelsky 1992; Clarke 1988; Ertürk 1998; Laidler 1999; Marcuzzo 2002; Hirai 2007). The present article contributes to that literature by examining Keynes’s “long struggle” over the notion of forced saving and how it is related to his expositions in The General Theory.

I argue that his new expositions in The General Theory are closely related to his intense intellectual debates with Dennis Robertson as well

History of Political Economy 48:3 DOI 10.1215/00182702-3638667 Copyright 2016 by Duke University Press

Correspondence may be addressed to Ho-Po Crystal Wong, Department of Economics, National Tsing Hua University R513, CTM, No.101, Section 2, Kuang-Fu Road, Hsinchu, Taiwan; e-mail: [email protected]. I am very grateful to Joe Chan, who brought the variorum drafts of chapter 23, which was removed from the later revisions of the Treatise on Money, to my atten- tion. I also thank him for his inspiration on the topic. I thank Simon Bilo, Thomas Cate, Harald Hagemann, Kevin D. Hoover, D. E. Moggridge, Gary Moon-cheung Shiu, participants at the 2012 History of Economics Society conference, and two anonymous referees for their very help- ful comments and suggestions. Citations to Keynes’s Collected Writings are to JMK, followed by the volume and page numbers.

516 History of Political Economy 48:3 (2016)

as to the criticisms of A Treatise on Money made by Ralph Hawtrey, who highlighted the flaws of using price as an equilibrating mechanism in the Treatise. Although all three—Keynes, Robertson, and Hawtrey—sub- scribed to fundamentally different monetary theories, all of their works, as David Laidler (1999) suggests, provide important visions of how in a world with money, coordination failures of investment and saving would lead to credit cycles. The exchange of ideas among Keynes, Robertson, and Hawtrey had been fruitful and significant to the development of each of their works, despite their disagreements on many critical points. As Austin Robinson ([1946] 1964, 51) affirmed Robertson’s influence on Keynes’s thought, “The important thing is that in the process of writing [his Banking Policy and the Price Level, Robertson] had forced not only himself, but also Keynes, to reexamine a number of fundamentals of monetary theory.”

This article illuminates the nature of the revisions of the Treatise in relation to the forced saving thesis and the series of modifications of the Fundamental Equations, which eventually resulted in the new formulation in his General Theory. D. E. Moggridge (1992, 551) suggested that “Keynes’s passage from his Treatise on Money to his General Theory has spawned a large literature. One of the matters this literature has not exten- sively considered is the way in which Keynes worked out his new ideas.” This article examines precisely that: how Keynes reached his new formu- lation in The General Theory. One relatively recent work on this subject is by Cristina Marcuzzo (2002). She studied the evolution of Keynes’s the- ory from the Treatise to The General Theory in the context of Richard Kahn’s influence on Keynes’s ideas (see Harcourt 1994 for more on Kahn’s influence on Keynes). The present article complements Marcuz- zo’s work by examining the evolution of Keynes’s monetary thought in the context of the link between Keynes’s disposing of the forced saving notion and his adopting new arguments, in which Kahn played a collaborative role, as Marcuzzo suggests. Kahn himself (1978, 548) also marked the connection between Keynes’s rejection of forced saving and “subsequent rapid progress towards a completely new formulation.” Yet the transition itself was not analyzed. My examination provides important insights on the fundamental differences and inherent incompatibility between Rob- ertson’s forced saving thesis and the later works of Keynes.

The Robertson-Keynes debate has been widely studied in the literature. John R. Presley (1979) provided an in-depth criticism of the forced saving thesis and the background to the debate. He emphasized the methodologi-

Wong / Keynes’s Departure from Forced Saving 517

1. In relation to the Robertson-Keynes debate, Fletcher (2000) provided a detailed account of Robertson’s failure to accept Keynes’s new method in a wider context.

cal differences between the two but not Keynes’s own changing position regarding forced saving and how it is related to the subsequent develop- ment of his theory. Gordon Fletcher (2000, 2008) provided a systematic account of the insights, parallels, and differences between Keynes’s works and Robertson’s works and how they are related to the later development of Keynes’s works.1 Eleonora Sanfilippo (2005) considered the evolution of the relationship between Keynes and Robertson from their intense col- laboration to their theoretical controversy that caused them to drift apart. Lilia Costabile (1997) related the undervaluation of Robertson’s works in the later development of monetary economics as an outcome of the Rob- ertson-Keynes debate. In contrast to these works, I look at this debate from a new angle and analyze it in detail from Keynes’s perspective, his struggle with the forced saving doctrine, which caused him to ultimately dispense with the forced saving notion in his monetary theory. The aboli- tion in turn allowed for a major breakthrough in his line of thought.

In addition, I link Ralph Hawtrey’s influence on Keynes to this depar- ture, and associate Keyens’s past struggle with the forced saving notion with his subsequent neglect of production time lags in his new monetary thought. Toshiaki Hirai (2007) also discussed Hawtrey and the Cam- bridge Circus’s influence on Keynes’s theoretical development, yet the attention was not on the link between their criticisms and his abolition of the forced saving notion.

To provide a detailed historical account of the evolution of Keynes’s monetary thought, in this essay I produce a chronology of Keynes’s gradual departure from the forced saving notion to the formulation of his theory of effective demand. A careful study of the evolution of Keynes’s line of thought sheds important light on the transformation of his ideas and allows us to understand better why The General Theory takes its current form.

The Root of the Forced Saving Doctrine

The doctrine of forced saving or similar terms used to describe the idea such as Jeremy Bentham’s ([1816] 1839) “forced frugality,” John Stuart Mill’s (1875) “forced accumulation,” or “automatic lacking” as termed by Dennis H. Robertson (1933), has appeared in various schools of eco- nomic thought, from the British classical school to the Austrian school.

518 History of Political Economy 48:3 (2016)

2. See chapter 4 (pt. 2) of Presley 1979 for an excellent outline of Robertson’s forced saving thesis.

3. Or when the velocity of the circulation of money increases. But the present article focuses on how an increase in the money supply (i.e., the action taken by the central bank) creates forced saving and results in a transfer of resources.

4. Under Robertson’s framework, capital consists of fixed capital and circulating capital (or working capital), and he argued that forced saving creates additional circulating capital, as the “expansion of Fixed Capital in a modern community comes about mainly (though not entirely) through the performance of Spontaneous Long Lacking by individuals and corporations [which include entrepreneurs reinvesting their profits in their business and the supply of funds by inves- tors through the new issues of securities]” (Robertson [1926] 1989, 50).

In general, it refers to a mechanism through which capital for the pro- ductive sector is created by increasing the money supply or by expanding the amount of credit extended by banks; such an increase or expansion induces a redistribution of real resources in favor of entrepreneurs and thus enhances employment. The doctrine can be dated back to Jeremy Bentham ([1816] 1839), who wrote on the effect of a rise in the money supply and described it as an “unprofitable income tax upon the income of fixed incomists” (45). He argued that increasing the money supply would enhance national wealth but at the expense of national comfort and justice.

Dennis H. Robertson and Forced Saving

Dennis H. Robertson (1933) further conceptualized the relation between forced saving and industrial fluctuations by adopting a Marshallian “day” analysis and the classical advanced wage-fund theory. The influence of Robertson’s ideas on saving and hoarding on Keynes’s works was substan- tial. As Keynes put it in the preface to the Treatise, “Mr. D.H. Robertson has cast a penetrating light on certain fundamental matters, and this book would never have taken its shape without the help of his ideas” (JMK, 5:xviii). Robertson’s forced saving thesis can be summarized as follows.2 There is a time lag between the payment of wages and the realization of output. The advanced wage fund going to the workers at time t – 1 (“the previous day”) is to be spent on consumption goods to be realized at time t (“any particular day”). Forced saving occurs when banks expand credit between t – 1 and t so that entrepreneurs receiving the extra credit from banks can bid more resources away from consumers.3 This action would drive up the price level and results in the creation of additional working capital at the expense of consumption.4 Workers’ consumption at time t would then fall short of the value of the income at time t – 1. Resources are then said to be transferred involuntarily from wage earners to producers

Wong / Keynes’s Departure from Forced Saving 519

through forced saving. Robertson ([1926] 1989, 47) calls this “automatic lacking,” which represents a process in which the “additional stream of money secures a part of goods for those from whom the additional stream of money flows and deprives the residue of the public of consumption which they would otherwise have enjoyed.”

Under the classical wage-fund theory, the sustenance of laborers comes from the accumulated fund of working capital. Scott Gordon (1973, 17) argued that this “led to the general conclusion that it is the accumulated fund of capital (or ‘circulating capital’ in some versions) that, by being ‘advanced’ to the laborers by the capitalists or employers, constitutes the fund out of which labor is paid.” Additional working capital and consequently more out- put and employment are created when resources are diverted from consump- tion to working capital, as the quantity of money increases.

In addition, what is known as “induced” lacking occurs when individuals seek to maintain the real value of their holdings of money by additional sav- ing as a result of automatic lacking. Lacking—and here Robertson means lacking of any kind—allows producers to increase circulating capital, including the employment of new workers, which is the foundation of Rob- ertson’s thesis. The banking system can create credit by way of loans. The credit expansion would result in more resources being diverted to the pro- duction of capital goods and is associated with inflation. Increasing the money supply which contributes to forced saving is sometimes “beneficial” to the community. For instance, in times when there is real productivity growth, and if the banks fail to take “appropriate action” (i.e., increase loans), the saving will “go to waste” and be “dissipated in the form of lower prices and increased consumption by the public at large.” This rests on the assumption that the workers hired (who are at the same time consumers) will not consume until the bank policy has realized its effect. “The bank can prevent this undesired result if, and only if, it makes new loans or invest- ments on a sufficient scale to prevent prices from falling.” This would enable the depositors’ “thrifty intentions to bear fruit” (Robertson [1940] 1956, 44).

Forced Saving and the Advanced Wage-Fund Doctrine

Of particular importance is the link between forced saving and the time lag in production. Forced saving relies upon a time lag between when workers get paid and the realization of output, without which there would be no forced saving to speak of: when the supply of loans goes up as a result of a credit expansion, the nominal payment to workers would also

520 History of Political Economy 48:3 (2016)

5. The Austrian school highlights the importance of the roundaboutness of production in the trade cycle (Böhm-Bawerk 1884; Mises [1924] 1934; Hayek [1929] 1933); Hawtrey particularly

rise immediately. As such, there would be no automatic lacking or dis- lacking, as the real consumption of individuals would never be reduced to a level below what they intend to spend. This entire thesis is built on the assumption that the payment to labor precedes output. This is widely known as the advanced wage-fund doctrine. John Stuart Mill well sum- marized it in the Fortnightly Review in May 1869, which later appeared in his Dissertations and Discussions (1875, 48) as “a sum of wealth, which is unconditionally devoted to the payment of the wages of labour. This sum is not regarded as unalterable, for it is augmented by saving, and increases with the progress of wealth; but it is reasoned upon as at any given moment a predetermined amount.” Although the wage fund is not rigidly fixed, its amount is strongly related to the working capital stock available to an economy. In classical political economy, labor is conceived to be paid out of a previously accumulated stock. To put it in the context of contempo- rary economics, the advanced wage fund resembles very much a rigid wage contract. And so conceptually the wage fund is set aside as a more or less fixed sum in advance, and workers get paid as output is realized, but their nominal wages, which have been fixed by the wage fund, would not be perfectly adjusted for the increase in the price level. In Keynes’s words, forced saving “raises prices in excess of any increase in the rates of money enjoyed by the factors of production as a whole; so that there is in this case a sacrifice of real income by the factors of production previously employed, the equivalent of which sacrifice thus becomes available for additional employment” (JMK, 13:107–8). However, the doctrines of forced saving and the advanced wage fund are in fact controversial under the classical regime. Forced saving is inconsistent with the classical assumption of full employment (see Gordon 1973). Frank Taussig (1894, 4) noted that the wage-fund theory was “copied from book to book” in a “rough and ill-defined form,” while Alfred Marshall (1888, 218) regarded the doctrine as “pretentious and misleading.”

Time Lag in Production and Forced Saving in the Treatise

With time lags in production, the role of the flow of capital in different stages of production could differ in the trade cycle.5 Some interesting

Wong / Keynes’s Departure from Forced Saving 521

paid attention to the role of stocks and inventories in the trade cycle, but he also acknowledged that “the construction of fixed capital in most cases is a lengthy process, and large commitments are therefore specially apt to overstrain credit when the turning point from expansion to con- traction arrives” (Hawtrey 1919, 109).

6. Moggridge (1992, 552) suggested that “one can regard his tables of contents as sketch maps of how he intuitively thought of the route through the larger problem in book form.” He pointed out that Keynes had a habit of structuring his arguments as a whole before he formally worked them out, and the drafts of his tables of contents for his works reflected the general scheme of his thought. Carefully examining the changes he made in these drafts is helpful in understand- ing the development of his ideas.

7. Working capital is defined as the aggregate of goods used in the course of production, manu- facture, transport, and retailing, as are required to avoid interruptions in the production process.

observations can be made from the draft tables of contents of the Treatise in regard to Keynes’s position on the roles played by different forms of capital in the course of credit cycles.6 Keynes defined capital in three major forms in the Treatise, namely, fixed, working, and liquid capital. In the draft table of contents dated June 13, 1925, chapters 10 and 11 in book 2 (“The Theory of Credit”) were titled “The Function of the Banks in Relation to the Supply of Working Capital” and “The Meaning and Sig- nificance of Working Capital,” respectively (JMK, 13:41–42). This shows clearly that in the earlier drafts of the Treatise, in the spirit of Robertson’s works, working capital has a particularly crucial role to play in the credit cycle. However, these chapters are no longer found in the subsequent draft tables of contents and in the finalized table of contents of the Treatise.

In the finalized version of the Treatise, the discussion on capital theory was treated as a digression and was included because “the fluctuations in the rate of investment have not been treated, sufficiently for my purpose, elsewhere” (JMK, 6:85). It can be inferred from this treatment that Keynes did not himself feel totally comfortable with his discussion of capital the- ory, as he had not established a solid theoretical foundation for the causal mechanism between credit inflation, the flow of capital, in particular working capital, and output.

It is nonetheless important to examine what appears in this digression to assess to what extent these capital theories of the credit cycle stand in contrast to his monetary thought. In the digression, he suggested that the three forms of capital are related to the credit cycle by the way they affect the investment rate, and they play somewhat different roles in the course of the trade cycle. The time-lag argument is implicit in the discussion. Among the three, working capital plays the most central role in the credit cycle.7 The longer the length of the production process and the higher the

522 History of Political Economy 48:3 (2016)

8. This term is used by Keynes to describe workers hired in the previous period (t – 1) under the advanced wage-fund theory. Noticeably, the output produced by the previously employed is realized in the next period (t), and the capital goods they produced would determine employ- ment in the next period (see JMK, 6:108).

intensity of employment, the more working capital is needed to avoid an interruption of the production process.

In the working capital theory, the banking system plays an important role in transferring real resources from unproductive consumption to working capital in times when trade is brisk, where unproductive con- sumption is “consumption which could be forgone by the consumer with- out reacting on the amount of his productive effort” (JMK, 6:111). Basi- cally, unproductive consumption is the consumption by the class that is not in the labor force and would not affect production. Banks should con- duct policies that ensure that the “wages fund” (Keynes’s term) can smoothly sustain the production process so as not to slow down trade. “It is the flow of income, available for consumption by the factors of produc- tion, which constitutes the true wages fund; and it is the distribution of this fund between relatively productive and relatively unproductive con- sumption which determines the volume of employment and of output” (JMK, 6:115).

As the production process lengthens, the forced saving role played by banks becomes increasingly crucial in maintaining the stability of trades. This is noticeably Robertson’s central thesis, in which expansion of output requires an increase in the provision of circulating or working capital, without which there cannot be changes in output and employment (Presley 1979). And for forced saving to produce effects, it has to change the wage fund available for production in the next period. Yet Keynes was actually doubtful about this doctrine of the “wages fund” and vaguely described it as follows: “In Marshall’s words, ‘it suggests a correlation between the stock of capital and the flow of wages, instead of a true correlation between the flow of the products of labour aided by capital and the flow of wages.’ But it has proved injurious to clearness of thought to demolish this doctrine without putting anything into its place” (JMK, 6:115).

In other words, the amount of working capital produced in the current period is closely related to employment in the next period. Figure 1 pro- vides a graphical summary of Keynes’s interpretation of Robertson’s forced saving framework. Keynes made a more careful distinction between the “currently employed” and the “previously employed.” 8 Noticeably, the

Wong / Keynes’s Departure from Forced Saving 523

previously employed were not necessarily workers who were employed in the next period. For instance, if unemployment increases, some of the pre- viously employed would then be unemployed. The level of employment in the next period depends on the wage fund available in period t, which is determined by the amount of working capital available. Forced saving transfers real resources from the production of consumption goods to working capital and this results in a larger wage fund available for produc- tion and employment in the next period. This mechanism of transference of resources by means of forced saving proposed by Keynes can be supple- mented by the following discussion in a letter to Robertson of May 28, 1925: “There is also the valid argument that a transference of consumption of such a kind that productive consumption (in my sense of the word) is substituted for unproductive consumption does increase saving, and there- fore short lacking [by which Keynes meant lacking used as finance for working capital]. If you transfer purchasing power from retired widows to the unemployed, on condition that the unemployed work, then there is increased short lacking” (JMK, 13:35).

Figure 1 Keynes’s interpretation of the forced saving mechanism

524 History of Political Economy 48:3 (2016)

9. Robertson’s hoarding in general refers to the act of holding money. Hoarding occurs when an individual holds money and “he is from his own point of view saving, but is taking no steps to ensure the creation of Capital. Unless others take such steps, the effect of his action . . . is that the consumption of other persons is increased by as much as his own consumption is dimin- ished” (Robertson [1926] 1989, 46). And the bank can transform hoarding into new circulating capital through lending: “It is almost equally tempting to suppose that the new Circulating Capital is the product of the New Hoarding of the owners of these deposits” (51–52).

Keynes’s Struggle over the Notion of Forced Saving in the Treatise

This capital theory as a digression in the Treatise, and Keynes’s dissatis- faction with the wage fund, reflect his hesitant position on the notion of time lags in production and forced saving as components of his monetary theory in the Treatise. Don Patinkin (1976, 28) points out that “the Trea- tise was originally conceived only as a systemization and elaboration of the theory that lay behind the policy recommendations of the Tract [on Monetary Reform], which had not been systematically developed there.” Yet Keynes’s drastic revisions of the Treatise reflect his uncertain state of mind. His exchanges with Robertson in the course of preparing the Trea- tise had in fact led him to further challenge the forced saving notion.

One question that seems to have constantly puzzled Keynes is, What exactly is forced saving? Keynes was dissatisfied with Robertson’s Bank- ing Policy and the Price Level ([1926] 1989), whose theory is built upon the forced saving notion. He agreed that credit inflation in general would produce some amount of new hoarding, but it was impossible to know the exact amount. Also, such new hoarding might be wholly or partially at the expense of other new investments.9 As such, the effect of forced saving in diverting resources to production was in doubt (JMK, 13:38–39). In spite of the disagreement, their discussions had a profound effect on each oth- er’s work. Robertson ([1926] 1989, 5) himself stated in the introductory chapter that “I have had so many discussions with Mr. J. M. Keynes on the subject matter of chapters V and VI, and have rewritten them so drasti- cally at his suggestion, that I think neither of us now knows how much of the ideas therein contained is his and how much is mine.” In contrast to Keynes, Robertson had consistently adhered to the forced saving doctrine. His criticisms of Keynes’s monetary thought in the Treatise and later of the formulation in The General Theory were closely tied to Keynes’s departure from the forced saving thesis.

In a letter to Keynes of February 27, 1925, Robertson expressed his discomfort with Keynes’s not adhering to the advanced wage-fund doc-

Wong / Keynes’s Departure from Forced Saving 525

10. A variorum of the drafts of this chapter is documented in JMK, 13:83–113.

trine. He stated explicitly that he was uncomfortable at the inclusion of all the earnings of fixed capital in the “wages bill,” as “the essence of the wages of labor is that they are advanced: while profit is only received at the time of sale” (JMK, 13:25–26). Such treatment had essentially invali- dated Robertson’s forced saving thesis.

Keynes’s skepticism over the notion of forced saving in preparation for the Treatise can be further elucidated by the removal of chapter 23 of the table of contents of August 1929, titled “The Part Played by the Banking System,” which hinted at banks transferring resources from the previously employed to the currently employed by means of a mechanism similar, if not identical, to forced saving.10 This chapter appears in the draft table of contents dated August 2, 1929. Presley (1979, 79) also noticed the removal of this chapter from the final version of the Treatise and “Keynes’ neglect of the forced saving doctrine which they [Keynes and Robertson] had put forward together in 1926.” He briefly related it to Keynes’s change of mind in terms of the differences in the definitions of saving and investment and in their relationship through trade cycles. In the following I will comple- ment Presley’s discussion by providing a detailed examination of this chapter so as to better understand why Keynes eventually removed it from the Treatise and how it is related to his gradual divergence from Robert- son’s forced saving thesis.

The arguments made by Keynes in the chapter are very much akin to those made by Robertson. A substantial part of the chapter discusses the importance of “working capital,” whose additional supply when business fluctuates is obtained mainly through the banks. The viewpoint in this chapter is that banks play a central role in adjusting the volume of working capital in the credit cycle. In section e of the chapter, “By Forced Transfer- ences of Purchasing Power from ‘Unproductive’ Consumers,” Keynes provided a discussion of credit inflation. He raised a question: “Is this [credit inflation] no better than a confession of failure, or does it do some- thing, though at the expense of price stability, really to satisfy the demand for additional investments?” And he added that “in doing this [i.e., in inflating credit], really allow some additional investment to take place” (JMK, 13:104).

What Keynes was unsure of was which individuals—depositors or the previously employed—have their wealth transferred to production by means of forced saving. This is related to the fact that he could never

526 History of Political Economy 48:3 (2016)

conceptually separate forced saving from conventional saving in his for- mulation. In fact, in his formulations in both the Treatise and The General Theory, saving was always treated as an aggregate. He questioned the validity of forced saving by carefully elaborating “at whose expense has this augmentation [of working capital] taken place.” He made a careful distinction between the previously employed and the depositors. In his view, the transference comes from the former but not the latter. “So long as depositors as a body are not drawing on their previous deposits for pur- poses of consumption, it is not their deposits (or the equivalent of these in available income) which they are consuming, but their current income” (JMK, 13:106).

Depositors typically save part of their income in each period, but it is usually their current income that determines the current consumption (or consumption in the next period in the advanced wage-fund theory). There- fore, despite the fact that increased prices diminish the real value of deposits, no additional real resources will be diverted from the depositors to the borrowers. The only group of individuals who could ultimately con- tribute to the transfer of real resources is the previously employed, who by assumption do not spend their income until the next period, when prices increase. This is the class that produced the output in the current period (see figure 1). In his framework, there is a trade-off between the supply of working capital and consumption goods, and the latter is consumed by the previously employed. And if the fund of working capital is augmented by increased credit from banks, the real income of the previously employed would fall and thus they would consume less in real terms. As Keynes suggested, “There is in this case a sacrifice of real income by the factors of production previously employed, the equivalent of which sacrifice thus becomes available to provide for additional employment” (JMK, 13:108). This chapter in the end was taken out of the final version of the Treatise.

In the following I shall provide some speculation as to why Keynes removed this chapter. I link this decision to his dissatisfaction with the application of the notion of forced saving in the explanation of the trade cycle and its inherent incompatibility with his Fundamental Equations.

First, it is unclear that the forced saving mechanism is compatible with Keynes’s pure theory of money in the Treatise. Recall in the Trea- tise that the second Fundamental Equation is given by (reproduced from JMK, 5:124)

П = W1 + I —– O

S,

Wong / Keynes’s Departure from Forced Saving 527

where П stands for the price level of output as a whole; W1 represents the rate of earnings per unit of output; I is investment; O is total output of goods; and S is saving. The profit equation is

Q = I – S,

where Q is total profit. For total profit to be positive, which creates a ten- dency for prices to go up and entrepreneurs to expand output, investment has to be larger than saving. The question here is whether price inflation and output expansion by means of the forced saving mechanism can be reconciled with the above equations without ambiguity.

Applying Robertson’s forced saving thesis to the Fundamental Equa- tions, credit expansion generates lacking, and overall saving would increase. It is uncertain how the increase in saving here would drive up investment to a level that would increase profit (and the price level), which creates forces for output to expand. Note that at this stage, Keynes had not yet recognized saving to depend on the level of output, unlike in his General Theory (see Davis 1980).

Patinkin (1976) presented evidence that the Fundamental Equations were developed after the draft table of contents of August 2, 1929. Remarkably, this was after the above “chapter 23” had been removed. “For the distinctive features of these equations is their dependence on the difference between savings and investment, and the question of this dif- ference does not seem to have been referred to until the draft table of contents of October 1928” (28–29). This is consistent with the above con- jecture that Keynes felt that the forced saving notion was inconsistent with his major analytical tools in the Treatise: the Fundamental Equations.

In addition, the notion of forced saving is founded on the supply-side argument where saving provides the finance for investment. As pointed out by Eprime Eshag (1963, 59), the doctrine of forced saving implicitly assumes full employment. Under this doctrine, an increased level of sav- ing would enlarge investment. It is based on the notion that through expanding credit, the banking system could exert a “forced” amount of saving (or real resources) from the consumers or, to be precise, the “previ- ously employed,” to be used in production and thus increase output in the future. In fact, it is a story of redistribution of resources between con- sumption and capital goods. The trade cycle can be adjusted through credit expansion that diverts resources to production (in Robertson’s term, an increase in “shortlacking” through credit expansion) in times of slumps. The question is whether this thesis, which is founded on full

528 History of Political Economy 48:3 (2016)

11. These are deposits that come from income created by loans through, for example, credit expansion and therefore their creation is a result of the increase in working capital.

12. These deposits concern mainly people hoarding money in times of uncertainty. So under uncertainty, savings deposits tend to increase.

employment, remains valid in times of unemployment. Presley (1979) reported Keynes’s criticisms of Robertson’s forced saving thesis, in which he questioned whether prices of consumption goods would rise in the presence of unused resources. As such, there can be no forced saving to speak of.

Apart from the inconsistency in having the full employment assump- tion in a model that ultimately seeks to explain the variation in output, Keynes contested the magnitude of real resources brought to production by means of forced saving. In response to Robertson’s draft of Banking Policy and the Price Level in 1925, he agreed that credit inflation tends to result in some amount of new hoarding, yet it is impossible to determine its amount a priori. Some of the so-called new hoarding may be created partly or wholly at the expense of other new investment. As such, the effectiveness of credit inflation is in doubt (JMK, 13:38–39).

Keynes also commented in the Treatise that aside from forced saving or automatic lacking, “there are other ways of augmenting the income depos- its,—for example, by a transfer from the savings deposits or by refraining from the purchase of securities with normal current savings;11 or the velocity of the income deposits, instead of their amount, may be increased” (JMK, 5:269).12 Interestingly, at this stage, Keynes still incorporated the idea of forced saving in his policy views. In 1929 Keynes published a pamphlet titled Can Lloyd George Do It? with Hubert Henderson in his advocacy of an expansion program. He mentioned that one source of new investment comes from “savings which now run to waste through lack of adequate credit” (Keynes and Henderson 1929, 35).

Keynes might appear obscure in his position regarding the forced saving notion in this period. Indeed, he described his hesitant state of mind in the preface to the Treatise as a “long struggle.” The preface to foreign editions of the Treatise written in 1932 described the notion of forced saving as

not the theory of Book III below, though it may have certain affinities with it. I believe that the ideas, which I have just endeavoured to sketch, are essentially unclear and that, if a thorough attempt is made to render them clear, they will undergo a series of modifications which will grad- ually have the effect of bringing them into conformity with my own

Wong / Keynes’s Departure from Forced Saving 529

theory. . . . For whilst I hold that the policy of the banking system influ- ences the difference between saving and the value of investment, I do not hold that there is any direct, necessary or invariable relationship between this difference and the amount of credit, . . . which could be deduced from a knowledge, however complete, of banking and cur- rency statistics. (JMK, 5:xxiv)

In my view, he was then unsure about the possible way to align the notion of forced saving with his theory of money in the Treatise. Also, he did not think the forced saving thesis is always empirically true, which made him cast further doubt on applying the thesis to the explanation of output fluc- tuation in the real world.

In the original preface to the Treatise, he confessed that the ideas in the book might appear confusing due to his struggle of thought:

I am acutely conscious of its defects. It has occupied me for several years, not free from other occupations, during which my ideas have been developing and changing, with the result that its parts are not all entirely harmonious with one another. . . . There is a good deal in this book which represents the process of getting rid of the ideas which I used to have and of finding my way to those which I now have. . . . I am not clear exactly what the difference is on this definition between “vol- untary” saving and “forced” saving. (JMK, 5:xvii–xix)

One unresolved question that remained after the publication of the Treatise is whether Robertson’s “theory of the causal relation between inflation and new hoarding” is in harmony with his Fundamental Equa- tions (JMK, 13:38). F. A. Hayek (1932, 133) actually interpreted Keynes’s monetary theory in the Treatise as a rejection of the terminology of forced saving and stated that Keynes replaced the notion with that of “investment being in excess of saving; and there is much to be said in favor of this.”

Toward The General Theory

In the subsequent development of Keynes’s works upon the publication of the Treatise, he became increasingly critical of the conceptual separa- tion of voluntary and forced saving and by 1934 at the latest he appeared to have completely abolished forced saving in his monetary theory of production.

Yet, the rejection of the forced saving notion was a gradual one. The transition can be best examined through Keynes’s correspondence with

530 History of Political Economy 48:3 (2016)

13. By this, Keynes meant his new definitions for income and saving.

Dennis Robertson in the early 1930s and the notes on Keynes’s lectures from 1932 to 1935, which were taken by his students and edited by Thomas K. Rymes (1989). Keynes’s departure from the forced saving notion can be roughly broken down into three phases after the publication of the Treatise.

Phase 1 (1929–32): Continued Struggle over Forced Saving

After the publication of the Treatise, Keynes attempted to clarify the thoughts that he had “put inadequately in Book Three” in a long letter written to Robertson dated March 22, 1932, and titled “Notes on the Defi- nition of Saving” (JMK, 13:275). He had clearly refined his interpreta- tion of the role of saving in business cycles: “S [conventional saving] is not the voluntary result of virtuous decisions. In fact, S′ is no longer the dog, which common sense believes it to be, but the tail. Not that the virtuous decisions are of no effect. But what they settle is, not the amount of sav- ing, but the relation between the consumption price level and the cost of production,—which is not, I think, what ‘common sense’ supposes” (JMK, 13:276). He further challenged the notion of forced saving and raised the following questions:

(1) Does “forced” saving arise only (a) when the banking system lends money to the government or to entrepreneurs or (b) when the velocity of circulation is increased? (2) Can it be regarded as a function of the amount by which prices rise or the amount by which deposits increase or the change in the velocity of circulation? (3) How, in exact terms, is this amount of forced saving to be measured? (JMK, 13:285)

Keynes thought that there were no satisfactory answers to the above ques- tions and concluded that in his Treatise

my critical departure from the previous theories of “forced” saving or “induced” and “automatic” lacking lies not in my definitions13 . . . but in my conclusion . . . that induced and automatic lacking are not simple functions of monetary factors such as the quantity of money or its velocity of circulation . . . , that they can arise in other ways than as a

Wong / Keynes’s Departure from Forced Saving 531

14. See JMK, 13:275–79, for the redefinition. 15. At this stage, Keynes’s monetary theory of production is still limited to a tendency anal-

ysis, so here the output level and the cost of production are fixed at the original level; ΔQ is the realized profit for the given output.

result (to quote D.H.R.) of “an additional daily stream of money being brought on to the market,” and that quantitatively, their amount cannot be deduced from banking statistics however complete. (JMK, 13:287)

In an attempt to present his ideas in the Treatise in a way that could be more easily understood, Keynes redefined income as E′ = E + Q to make it parallel to the conventional definition of income, which defines income as the cost of production plus profit.14 Keynes himself confessed in the preface to foreign editions of the Treatise that “my definition of income is thought paradoxical because I exclude from it (as explained below) wind- fall profits and losses, and my definition of saving, being the excess of income thus defined over expenditure on consumption, corresponds to my definition of income” (JMK, 5:xxiii). Keynes provided two alternative definitions for savings and income (reproduced from JMK, 13:275–77):

S = E – F and S′ = E′ – F. (1932.1)

E and E′ = E + Q. (1932.2)

The former definition of saving is the one that appeared in the Treatise, while the latter one corresponds to the conventional definition of saving and must be identical to the value of current investment. For example, Robertson defined S = E + Q – PR, where PR stands for the expenditure of the community on consumption goods (see JMK, 13:235). E is the cost of production; F denotes the amount of money spent on consumption. S is exactly the saving in the Treatise. E′ refers to the cost of production E plus the net profit of entrepreneurs Q in the Treatise, “which is the total income in Hawtrey’s, Hayek’s and D.H.R.’s sense, and in the sense to which I have now bowed the knee” (JMK, 13:275).

Importantly, in the notes Keynes placed his focus on F (consumption) rather than S, as he felt that it would “avoid the difficulties which many readers seem to have found in my definition of saving. This is possible because F is independent of the rival definitions of income and saving, since E – S = E′ – S′ = F ” (JMK, 13:277).

Evaluating at fixed output so that the cost of production does not change and ΔE = 0, this gives15

ΔQ = ΔI + ΔF, (1932.3)

532 History of Political Economy 48:3 (2016)

where Q is the net profit of entrepreneurs, I is investment, and F is the amount of money that is spent on consumption.

The above equation can be rewritten as

ΔI = ΔQ – ΔF. (1932.4)

This is a remarkable modification, as it allows Keynes to focus on investi- gating how investment affects profits (and later, how it affects output) and subsequently the relationship between investment and output, which in turn determines consumption and saving. This is essentially the skeleton of his General Theory.

An important point is that the new exposition is in fact compatible with the Fundamental Equations, yet the role of saving has been much down- played. This detachment from saving, which had played a central role in the Treatise, is parallel with his continuous struggle in the conception of savings and his difficulty in trying to model saving as a determinant of output.

However, evidence shows that Keynes had not yet completely abolished the forced saving notion at this stage. In the notes, he still attempted to put his formulation into the context of Robertson’s forced saving: “For D.H.R. has never given a hint that he means by ‘automatic lacking’ exactly the same as what I mean by Q. . . . I seek the interpretation by throwing my mind back to what I used to mean by ‘voluntary’ and ‘forced’ saving when I began my enquiries” (JMK, 13:283). His attempt to reconcile forced sav- ing with his formulation is similarly shown in the notes of his student dated October 24, 1932—seven months after the “Notes on the Definition of Saving” were sent to Robertson. At this point, Keynes still held some belief that the forced saving notion can be put in conformity with his the- ory and that the differences between his view and Robertson’s regarding saving were primarily terminological:

D + I + F = E′ = E + Q, (1932.5)

where D stands for disbursement, I denotes investment, and F denotes consumption.

Q = I – (E – F ). (1932.6)

Hence

Q = I – S (1932.7)

Wong / Keynes’s Departure from Forced Saving 533

and

S′ = I. (1932.8)

S′ is defined as “surplus income” by Keynes, which must be equal to investment.

S′ = S + Q. (1932.9)

Note that S here is the same as the saving defined in the Treatise. In the lecture notes, there are some important remarks for these formula- tions: “Saving is something that has to occur to make more investment possible . . . and S = S1 + S2 and S1, S2 and Q actually correspond to Dennis Robertson’s three kinds of lacking. . . . Forced saving may be considered as Q” (Rymes 1989, 61–62). Under the assumption that saving is the finance for investment, if S is voluntary saving and Q is positive such that I > S, Q would then be forced saving in the context of Robertson’s analysis.

Keynes apparently was not totally convinced by this argument, as he later on discarded the forced saving notion. Upon the abolition of the forced saving notion, saving was no longer treated as what makes invest- ment possible.

Phase 2 (1933): Intensified Doubt over the Validity of Forced Saving

A letter sent to Robertson dated May 20, 1933, indicates that Keynes was drifting away further from the forced saving doctrine: “I should have no objection to this use of the term ‘hoarding,’ once I had got used to it. The objection lies in the suggestion it conveys that it has something to do with actual holdings of cash. It is today, for the first time, that I discover I have been misled by this suggestion and that hoarding can occur during a period in which no individual alters his holding of cash” (JMK, 13:308). Apparently, Keynes increasingly found the period approach adopted by Robertson problematic. In December 1933, Keynes published a paper titled “Mr. Robertson on ‘Saving and Hoarding’” in the Economic Jour- nal; he argued that he was “not quite clear whether Mr. Robertson’s hoard- ing during successive periods can be aggregated over a complete period” and he admitted the flaws in his Treatise: “I do not now consider my anal- ysis in that book to be as logical as I can make it” (JMK, 13:330). Notice- ably, Keynes’s discussion with Robertson on subjects related to forced

534 History of Political Economy 48:3 (2016)

16. The Robertson-Keynes correspondence on savings did not resume until after the publi- cation of The General Theory. The first piece of evidence of the revival of discussions is a letter from Keynes dated September 20, 1936. The initiating letter from Robertson has, unfortunately, not survived (JMK, 14:86–87).

saving and the Treatise appeared to have discontinued after the publica- tion of this paper, and Keynes’s monetary thought continued to depart from Robertson’s thesis.16 Around the same period, he had further revised his Fundamental Equations, as indicated by the lecture notes dated November 13, 1933, taken by his students. In these notes,

Y = E + Q = C + I, (1933.1)

S = Y – C, (1933.2)

and

S = Y – C = E + Q – C, (1933.3)

where C is equivalent to F in his 1932 lecture and Y is aggregate disburse- ment. (1933.3) is the conventional definition for saving. And it is explicitly stated that “saving” in the Treatise will be denoted by S′ (which was previ- ously denoted by S) and equals E + Q′ – C, where Q′ represents the flow of quasi-rent relevant to long-period expectations, and he subsequently assumed it to be zero, which would make it equivalent to the definition in the Treatise.

The Fundamental Equations now become

ΔS = ΔQ + Δ(E – C ) = ΔI, (1933.4),

ΔS′ = Δ(E – C ), (1933.5)

and

ΔQ = ΔS – ΔS′ = ΔI – ΔS′. (1933.6)

The change in output will be driven by ΔQ. Firms will expand output if ΔQ is positive and vice versa. In equilibrium, ΔQ = 0. Keynes further pro- vided arguments in this lecture that Y is governed by I. So the above sug- gests that changes in I ultimately determine the change in Y; consumption and saving are endogenously determined, and both increase with Y. Clearly the backbone of The General Theory is in the making. The two alternative definitions of saving used in the new Fundamental Equations above represent Keynes’s attempt to integrate his theory in the Treatise into his new monetary theory of production. These are in fact products of

Wong / Keynes’s Departure from Forced Saving 535

17. See Patinkin 1993 for a detailed examination of the chronology of The General Theory.

his struggle with the conception of saving, which might appear confusing to many readers and to a large extent to Keynes himself too, as indicated by the subsequent drastic shift in his position on forced saving.

Indeed, the year 1933 appears to mark an important evolution in Keynes’s monetary thought. D. E. Moggridge also raised the point that Keynes had changed the contents of his lectures considerably from the previous year in the Michaelmas term of 1933 (JMK, 13:420). And in his biography of Keynes, he provided evidence that during that period Keynes was attempting to “make his break with traditional theory” (Moggridge 1992, 561). In line with this speculation, the first surviving complete table of contents of The General Theory was produced in December 1933 (JMK, 13:421–22). This is consistent with the idea that the major break- through in Keynes’s monetary theory occurred after his drifting away from the forced saving notion. This shift in his position on forced saving allowed Keynes to form a clearer view of the structures of his argument, as reflected by the production of his possibly first complete table of con- tents to The General Theory, or at the very least, the first draft of a table of contents that was worth keeping a good record of from Keynes’s view.

This breakthrough in Keynes’s monetary thought synchronizes with Don Patinkin’s (1976, 79) conclusion that “Keynes formulated his theory of effective demand during 1933, and in all probability during the first half of that year.” Robert Dimand (1988, 167) also argued that “Keynes made a decisive advance between his 1932 and 1933 lectures.”17 These conclusions suggest that the formulation of the theory of effective demand is closely connected with Keynes’s departure from the forced saving notion.

Phase 3 (1934 and Onward): The Ultimate Departure

Keynes became increasingly critical of the forced saving doctrine after 1933. In the lecture notes taken by his student dated November 19, 1934, Keynes made jokes about the forced saving notion and discussed how sav- ing becomes a curious concept when one has a curious concept of income, as in his Treatise. Furthermore, he pointed out that “Robertson gets a funny definition of savings because he defines today’s income as yester- day’s income. The term ‘forced saving’ is also nonsense because no one could ever explain what it means” (Rymes 1989, 168). This sarcastic

536 History of Political Economy 48:3 (2016)

18. Although some of the discussion above has been based on the lecture notes of his stu- dents (Rymes 1989) and arguably they only reflected what the students thought Keynes was saying and could potentially carry some inaccuracy in what he really meant, the notes clearly show that Keynes had become increasingly detached from the forced saving notion during the 1932–35 period. Also, the notes exhibit high levels of consistency and agreement over what had been said by Keynes (Moggridge 1992).

remark about Robertson’s thesis marks his ultimate abolition of the forced saving thesis in diverting resources to investment. Consistent with this line of development in his thought, Keynes made a similar comment in the lecture dated October 28, 1935: “Robertson in the Economic Journal, in 1933, defined income as equal to yesterday’s consumption and today’s investment. His saving equals income less today’s consumption, when saving exceeds investment, it just means that income is falling” (168).

To analyze this argument and link it with Keynes’s own theory of out- put, algebraically this means

Et = Ct – 1 + It (1935.1)

and

St = Et + Ct. (1935.2)

Therefore we have

It + Ct – 1 = St – Ct. (1935.3)

The above suggests that when St is larger than It, Ct must be less than Ct – 1. Based on Keynes’s production theory, C would only drop when income drops. Based on Keynes’s reframing of Robertson’s framework, the deri- vation indicates the following: when current saving is larger than current investment, current output would fall. From Keynes’s point of view, this result invalidates Robertson’s forced saving mechanism under his own framework and shows that they are logically incompatible. Based on the evidence presented in the present paper, in my view it is safe to conclude that by 1934 the forced saving notion was completely dispensed with in Keynes’s monetary theory.18

Ralph Hawtrey’s Influence on Keynes

With the abolition of the forced saving thesis, the next important question is, What new components did Keynes bring into his new formulation that are related to this past struggle with the forced saving notion? Patrick

Wong / Keynes’s Departure from Forced Saving 537

Deutscher (1990) suggested that the major influence Ralph Hawtrey had on Keynes in the development of The General Theory was from his criti- cism of the Treatise. For the purpose of the present article I restrict my discussion to the part of Hawtrey’s influence that contributed to Keynes’s abolition of the forced saving notion.

In the wake of the publication of the Treatise, Keynes had been engaged in extensive exchanges with Hawtrey. These exchanges turn out to have played an important role in the refinement of Keynes’s monetary theory. As Keynes himself admitted in his response to Hawtrey for his criticisms of the Treatise in a letter of February 16, 1931, “It is very seldom indeed that an author can expect to get as a criticism anything so tremendously useful to himself” (JMK, 29:10).

Hawtrey emphasized that Keynes’s Fundamental Equations in fact do not “exhibit the causal process by which the price level is determined” (JMK, 5:120): the divergence between saving and investment is merely a restatement of the divergence between prices and costs. Yet the ultimate cause of the divergence remains unclear.

Mr. Keynes’s formula only takes account of the reduction of prices in relation to costs, and does not recognize the possibility of a reduction of output being caused directly by a contraction of demand without an intervening fall of price. . . . The discrepancy between investment and saving is not the cause of the divergence between prices and costs; it is the divergence between prices and costs. The cause of the divergence (in so far as it is monetary) is to be found in a change in demand, i.e. in the consumers’ outlay. (JMK, 13:152–53)

Further, in a letter of December 6, 1930, Hawtrey emphasized the sig- nificant role played by change in demand in the trade cycle in the form of accumulation (or diminution) of unsold stocks. He argued that “the excess of saving over investment cannot be the cause of the reduction of output by entrepreneurs. The cause is to be found in the accumulating stocks which lead them to anticipate a future price fall. . . . The first result of a change in demand is a reaction upon sales, stocks and output, and only afterwards are prices affected” (JMK, 13:165–68).

Keynes apparently dealt with his criticism seriously. His reply to Haw- trey dated June 1, 1932, illustrates this point. “I now put less fundamental reliance on my conception of savings and substitute for it the conception of expenditure” (JMK, 13:172). Remarkably, this statement also suggests a close connection between his constant struggle with the conception of saving and his new emphasis on effective demand.

538 History of Political Economy 48:3 (2016)

19. Deutscher (1990) highlighted the differences in the mechanism that alters output between Keynes’s and Hawtrey’s formulations. The former relies on the “psychological law” of consumption, while the latter focuses on the excess of incomes over consumption that creates the tendency for capital outlay and thus future income to increase.

E. G. Davis (1980) pointed out that Hawtrey was the one who intro- duced to Keynes changes in output as a mechanism in equilibrating investment and saving and that this proposition was apparently well taken by Keynes. Patrick Deutscher (1990, 105), in contrast, did not fully accept this view but agreed that “what is firmly established is that Hawtrey’s criticisms were among the factors which pushed Keynes to an analysis incorporating quantity adjustments not proximately caused by price changes.”19 Toshiaki Hirai (2007, 288) also recognized Hawtrey’s influ- ence on Keynes in Hawtrey’s stressing the importance of effective demand, and Hawtrey’s criticisms “made a great contribution to weaning Keynes from the Treatise framework, and setting him off—after some hesitation—in the direction of the General Theory.”

One other noteworthy point is the connection of Hawtrey’s criticism with the abolition of the forced saving notion: Hawtrey’s proposed stages of adjustment in the trade cycle in which output adjustment precedes changes in prices are basically in the reverse order of Robertson’s. As sug- gested by Fletcher (2008), the means of adjustment in Robertson’s forced saving framework involved a “two-stage process” in which a change in demand affects price in stage 1, which makes forced saving possible and produces a subsequent change in output in stage 2.

In contrast to Robertson, Hawtrey, in a memorandum on the Treatise written for the Macmillan Committee in 1930 and reproduced and expanded in The Art of Central Banking (1932), convincingly brought for- ward the argument that in the trade cycle, adjustment in output precedes any changes in prices: “Now if anything occurs to affect the demand for goods of any kind or of all kinds, the first result is an increase or decrease in sales at existing prices, and therefore a decrease or increase in stocks of the goods concerned” (336–37). This invalidated the adjustment mecha- nism proposed in Keynes’s Fundamental Equations in the Treatise, in which the disparity between prices and costs (alternatively, between investment and saving) is the force that creates inducement for output to change. Based on Hawtrey’s argument, price disparity is the symptom rather than the cause of the trade cycle. What naturally follows from this argument is that the forced saving mechanism, which diverts resources from consumers to producers by means of credit inflation, cannot be a

Wong / Keynes’s Departure from Forced Saving 539

central part of the explanation of the trade cycle, for prices only adjust at the later stage of the trade cycle and output could fluctuate even without any change in price. This equilibrating mechanism of output as proposed by Hawtrey is highly serviceable in Keynes’s monetary thought, not only on the grounds of its realism but also because it allows him to get rid of the forced saving notion and end his long struggle.

Suppression of the Time Element in The General Theory

Keynes’s abandonment of the notion of forced saving in The General The- ory came along with the suppression of time in his new exposition. Ana- lytically, his formulation in The General Theory is static. This stands in stark contrast to Robertson’s works in which time had always played a central role: forced saving and the advanced wage-fund theory that war- rants it. From Robertson’s perspective, the oversimplification of static eco- nomic models might lead to false conclusions because the process of tran- sition is of prime importance in the trade cycle (Presley 1983). From Keynes’s perspective, however, the complication brought by the step-by- step approach of Robertson and the resulting decomposition of saving only added complexity to the problem but yielded no analytical signifi- cance. His neglect of time in The General Theory reflects this thought. The abolition of the forced saving notion and the advanced wage fund enables Keynes to do away with the theoretical complication in introduc- ing time in saving and production in his exposition. Keynes himself described it in The General Theory as follows: “To me at least it [the pro- cess by which equilibrium is reached in The General Theory] was a wholly new idea. It has nothing to do with velocities of circulation, time- lags, etc. though these things enter a detailed, formal analysis of the order of events” (JMK, 14:90).

Parallel to his treatment of saving, the time element in investment is similarly suppressed. In Robertson’s thesis, just as saving is divided into voluntary and involuntary forms, investment primarily takes the form of fixed and working capital. Each of them has a different role to play in the course of a trade cycle. In Keynes’s new formulation, investment appears only as an aggregate. Also, unlike the arguments made by Hawtrey or Robertson, in The General Theory Keynes no longer emphasized the time dimension in the production of different forms of capital or stocks and their different roles in the trade cycle. The segregation of capital goods

540 History of Political Economy 48:3 (2016)

that involve different lengths of the production process is only significant insofar as their expected lives would affect the profitability of the produc- tion of investment goods.

In defense of his new position, Keynes wrote a note titled “Notes on the Measure of ‘Roundaboutness’” and provided conditions under which an increase in capital is not associated with a lengthening of the production period, as most exponents on the subject purported, the Austrian school in particular. His derivation suggested that it is unclear how the “round- aboutness” or length of production would change with the same amount of capital while capital’s employment rate varies (JMK, 29:155–57). He wrote, “If, however, we are dealing with a community subject to change or one which is not in equilibrium, can any clear meaning be given to ‘the length of the production period’? If so, I do not know what it is. Perhaps those who apply this concept to credit cycle problems will instruct us as to what it means in this context” (157).

In The General Theory, capital goods of different natures are signifi- cant only in the sense that their different prospective yields formed by entrepreneurs’ expectations would affect the amount of capital to be pro- duced, which determines current investment in the aggregates. He wrote that “my theory of marginal efficiency must also be applied to working capital. Certainly it must, and I intend so to apply it. But I do not see what special considerations arise. Exactly the same treatment applies. The peculiarity both of working and of liquid capital seems to me to reside in the comparative shortness of its life rather than in any other peculiarity” (JMK, 13:585).

The suppression of time elements in his new expositions had certainly upset Robertson and Hawtrey, whose theses had been firmly founded on the time dimension in production. His new treatment nonetheless was closely related to their intensive discussion, which had made Keynes cog- nizant of the theoretical difficulty in incorporating time in his model. As Richard Kahn (1984, 125) pointed out in his Mattioli lectures, “Had Keynes given way [i.e., recognized the time element] he would have had largely to rewrite the book.” And the abolition of the forced saving notion had made this simplification more plausible.

The End of the Struggle: The New Arguments

In the Treatise, price is the equilibrating mechanism, as the quantity theory of money was deeply rooted in Keynes’s “habitual modes of thought”

Wong / Keynes’s Departure from Forced Saving 541

(JMK, 7:xxiii). Hawtrey made a convincing argument that in the trade cycle output adjustment precedes changes in prices. This point had been well taken by Keynes, and the equilibrating role of price adjustment had been replaced by changes in output in The General Theory, whereas the role of price, which played a central role in his Treatise, had been much downplayed in his new formulation. It naturally rendered the forced saving notion invalid in his monetary theory and ended Keynes’s long struggle.

In The General Theory, the notion of forced saving disappeared entirely in his argument and so did the role played by the variation in capital with different time lengths of production in the trade cycle. Keynes summa- rized these substantial changes made in The General Theory as follows: “The relation between this book and my Treatise on Money, which I pub- lished five years ago, is probably clearer to myself than it will be to others; and what in my own mind is a natural evolution in a line of thought which I have been pursuing for several years, may sometimes strike the reader as a confusing change of view” (JMK, 7:xxi–xxii).

Conclusion

The General Theory is a product of Keynes’s long struggle with the forced saving thesis and the time length in the production process associated with it. I have attempted to trace out the evolution of Keynes’s monetary thought primarily from the Treatise on Money through The General The- ory and have provided a careful examination into his abolition of the forced saving doctrine and how it is related to the development of his monetary thought, which he himself described as a “natural evolution in a line of thought.” I have presented evidence that the substantial but gradual changes in the theoretical constructions that constitute his General The- ory reflect his continuous effort to bring the notion of forced saving in conformity with his monetary theory, which eventually proved futile.

After his long struggle over the forced saving notion, he concluded that “no one could ever explain what it means” (Rymes 1989, 168), and in the end he chose to escape from it and took an alternative approach in his new formulation.

Starting from taking Robertson’s work as “at first sight a superb theory about fluctuations” (JMK, 6:1) to questioning the notion of forced saving, and to finally focusing on effective demand in his theory of production and employment, the criticisms from Robertson and Hawtrey and his exchanges with them played a crucial part in shaping Keynes’s ideas into

542 History of Political Economy 48:3 (2016)

what appears in The General Theory. Although the notion of forced sav- ing was eventually abandoned, the process of the struggle with the con- ception of saving doubtlessly constitutes an important part of Keynes’s intellectual development.

References

Amadeo, E. J. 1989. Keynes’s Principle of Effective Demand. Aldershot: Edward Elgar.

Bentham, J. (1816) 1839. The Works of Jeremy Bentham. Vol. 3. Edited by J. Bowring. Edinburgh: William Tait.

Böhm-Bawerk, E. 1884. Kapital und Kapitalzins. London: Macmillan. Clarke, P. 1988. The Keynesian Revolution in the Making, 1924–1936. New York:

Oxford University Press. Costabile, L. 1997. “Robertson and the Post-Keynesian Approach to Growth and

Cycles.” In Capital Controversy, Post-Keynesian Economics, and the History of Economic Thought, edited by P. Arestis, G. Palma, and M. Sawyer, 1:285–98. New York: Routledge.

Davis, E. G. 1980. “The Correspondence between R. G. Hawtrey and J. M. Keynes on the Treatise: The Genesis of Output Adjustment Models.” Canadian Journal of Economics 13 (4): 716–24.

Deutscher, P. 1990. R. G. Hawtrey and the Development of Macroeconomics. Ann Arbor: University of Michigan Press.

Dimand, R. W. 1988. The Origins of the Keynesian Revolution: The Development of Keynes’ Theory of Employment and Output. Aldershot: Edward Elgar.

Ertürk, K. A. 1998. “From the Treatise to the General Theory: The Transformation of Keynes’s Theory of Investment in Working Capital.” Cambridge Journal of Eco- nomics 22 (2): 173–85.

Eshag, E. 1963. From Marshall to Keynes: An Essay on the Monetary Theory of the Cambridge School. Oxford: Basil Blackwell.

Fletcher, G. 2000. Understanding Dennis Robertson: The Man and His Work. Chel- tenham: Edward Elgar.

———. 2008. Dennis Robertson. Hampshire: Palgrave Macmillan. Gordon, S. 1973. “The Wage-Fund Controversy: The Second Round.” History of

Political Economy 5 (1): 14–35. Harcourt, G. C. 1994. “Kahn and Keynes and the Making of The General Theory.”

Cambridge Journal of Economics 18 (1): 11–23. Hawtrey, R. G. 1919. Currency and Credit. London: Longmans Group. ———. 1932. The Art of Central Banking. London: Longmans Group. Hayek, F. A. (1929) 1933. Monetary Theory and the Trade Cycle. Translated by N.

Kaldor and H. Croome. London: Routledge & Kegan Paul. ———. 1932. “A Note on the Development of the Doctrine of ‘Forced Saving.’”

Quarterly Journal of Economics 47 (1): 123–33.

Wong / Keynes’s Departure from Forced Saving 543

Hirai, T. 2007. “How, and for How Long, Did Keynes Maintain the Treatise Theory?” Journal of the History of Economic Thought 29 (3): 283–307.

Kahn, R. 1978. “Some Aspects of the Development of Keynes’ Thought.” Journal of Economic Literature 16 (2): 545–59.

———. 1984. The Making of Keynes’ General Theory. Cambridge: Cambridge Uni- versity Press.

Keynes, J. M. (1930) 1971a. The Collected Writings of John Maynard Keynes. Vol. 5, A Treatise on Money: The Pure Theory of Money. Edited by E. Johnson and D. Moggridge. London: Macmillan.

———. (1930) 1971b. The Collected Writings of John Maynard Keynes. Vol. 6, A Treatise on Money: The Applied Theory of Money. Edited by E. Johnson and D. Moggridge. London: Macmillan.

———. (1936) 1973. The Collected Writings of John Maynard Keynes. Vol. 7, The General Theory of Employment, Interest, and Money. Edited by E. Johnson and D. Moggridge. London: Macmillan.

———. 1973a. The Collected Writings of John Maynard Keynes. Vol. 13, The Gen- eral Theory and After: Part I; Preparation. Edited by D. Moggridge. London: Macmillan.

———. 1973b. The Collected Writings of John Maynard Keynes. Vol. 14, The Gen- eral Theory and After: Part II; Defence and Development. Edited by D. Mog- gridge. London: Macmillan.

———. 1979. The Collected Writings of John Maynard Keynes. Vol. 29, The General Theory and After: A Supplement. Edited by D. Moggridge. London: Macmillan.

Keynes, J. M., and H. D. Henderson. 1929. Can Lloyd George Do It? An Examination of the Liberal Pledge. London: Nation and Athenaeum.

Laidler, D. 1999. Fabricating the Keynesian Revolution: Studies of the Inter-war Literature on Money, the Cycle, and Unemployment. Cambridge: Cambridge Uni- versity Press.

Marcuzzo, M. C. 2002. “The Collaboration between J. M. Keynes and R. F. Kahn from the Treatise to the General Theory.” History of Political Economy 34 (2): 421–47.

Marshall, A. 1888. “Wages and Profits.” Quarterly Journal of Economics 2 (2): 218–23. Meltzer, A. 1988. Keynes’ Monetary Thought: A Different Interpretation. New York:

Cambridge University Press. Mill, J. S. 1875. Dissertations and Discussions: Political, Philosophical, and His-

torical. Vol. 3. New York: Henry Holt. Mises, L. von (1924) 1934. The Theory of Money and Credit. Translated by H. E.

Batson. London: Jonathan Cape. Moggridge, D. 1992. Maynard Keynes: An Economist’s Biography. London: Routledge. Patinkin, D. 1976. Keynes’ Monetary Thought: A Study of Its Development. Durham,

N.C.: Duke University Press. ———. 1982. Anticipations of the General Theory? And Other Essays on Keynes.

Chicago: University of Chicago Press. ———. 1993. “On the Chronology of the General Theory.” Economic Journal 103

(418): 647–63.

544 History of Political Economy 48:3 (2016)

Presley, J. R. 1979. Robertsonian Economics: An Examination of the Work of Sir D. H. Robertson on Industrial Fluctuation. New York: Holmes and Meier Publishers.

———. 1983. “D. H. Robertson: Some Restoration.” Journal of Post Keynesian Eco- nomics 6 (2): 230–40.

Robertson, D. H. (1926) 1989. Banking Policy and the Price Level. New York: Augustus M. Kelley.

———. 1933. “Saving and Hoarding.” Economic Journal 43 (171): 399–413. ———. (1940) 1956. Essays in Monetary Theory. London: Staples Press. ———. 1959. Lectures on Economic Principles. Vol. 3. London: Staples Press. Robinson, E. A. G. (1946) 1964. “John Maynard Keynes, 1883–1946.” In Keynes’

General Theory: Reports of Three Decades, edited by R. Lekachman, 13–86. London: Macmillan.

Rymes, T. K. 1989. Keynes’s Lectures, 1932–35: Notes of a Representative Student. London: Macmillan.

Sanfilippo, E. 2005. “Keynes’s Valuable Opponent and Collaborator: The Correspon- dence between Keynes and Robertson.” In Economists in Cambridge: A Study through Their Correspondence, 1907–1946, edited by M. C. Marcuzzo and A. Rosselli, 58–74. New York: Routledge.

Skidelsky, R. 1992. John Maynard Keynes. Vol. 2, The Economist as Saviour, 1920– 1937. London: Macmillan.

Taussig, F. W. 1894. “The Wages-Fund Doctrine at the Hands of German Econo- mists.” Quarterly Journal of Economics 9 (1): 1–25.

Copyright of History of Political Economy is the property of Duke University Press and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.